Libraries

# Calculus
import pandas as pd

import os

# Data viz
import plotly.express as px
import plotly.io as pio
pio.renderers.default='notebook'

Brief

• numbers of crimes
  • per category
  • per location

Getting the data

columns_keep = [
#'division_number',
'date_reported',
'date_occurred',
#'area',
'area_name',
#'reporting_district',
#'part',
#'crime_code',
'crime_description',
#'modus_operandi',
'victim_age',
'victim_sex',
'victim_descent',
#'premise_code',
'premise_description',
#'weapon_code',
'weapon_description',
#'status',
'status_description',
#'crime_code_1',
#'crime_code_2',
#'crime_code_3',
#'crime_code_4',
'location',
#'cross_street',
'latitude',
'longitude',
]

# fetch all the data from the raw_data folder
current_dir = os.getcwd()
current_dir

file_path = os.path.join(current_dir, '..', 'raw_data', 'data.csv')
df = pd.read_csv(file_path)

df = df[columns_keep]
df['counter']=1

# Dates
df['date_occurred'] = pd.to_datetime(df['date_occurred'], errors='coerce')
df['year_occurred'] = df['date_occurred'].dt.year
df['month_occurred'] = df['date_occurred'].dt.month
df['hour_occurred'] = df['date_occurred'].dt.hour

df.head(3)

	date_reported	date_occurred	area_name	crime_description	victim_age	victim_sex	victim_descent	premise_description	weapon_description	status_description	location	latitude	longitude	counter	year_occurred	month_occurred	hour_occurred
0	2020-01-08	2020-01-08 22:30:00	Southwest	BATTERY - SIMPLE ASSAULT	36	F	B	SINGLE FAMILY DWELLING	STRONG-ARM (HANDS, FIST, FEET OR BODILY FORCE)	Adult Other	1100 W 39TH PL	34.0141	-118.2978	1	2020	1	22
1	2020-01-02	2020-01-01 03:30:00	Central	BATTERY - SIMPLE ASSAULT	25	M	H	SIDEWALK	UNKNOWN WEAPON/OTHER WEAPON	Invest Cont	700 S HILL ST	34.0459	-118.2545	1	2020	1	3
2	2020-04-14	2020-02-13 12:00:00	Central	SEX OFFENDER REGISTRANT OUT OF COMPLIANCE	0	X	X	POLICE FACILITY	NaN	Adult Arrest	200 E 6TH ST	34.0448	-118.2474	1	2020	2	12

Checking the data

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 852950 entries, 0 to 852949
Data columns (total 17 columns):
 #   Column               Non-Null Count   Dtype         
---  ------               --------------   -----         
 0   date_reported        852950 non-null  object        
 1   date_occurred        852950 non-null  datetime64[ns]
 2   area_name            852950 non-null  object        
 3   crime_description    852950 non-null  object        
 4   victim_age           852950 non-null  int64         
 5   victim_sex           740344 non-null  object        
 6   victim_descent       740336 non-null  object        
 7   premise_description  852432 non-null  object        
 8   weapon_description   296748 non-null  object        
 9   status_description   852950 non-null  object        
 10  location             852950 non-null  object        
 11  latitude             852950 non-null  float64       
 12  longitude            852950 non-null  float64       
 13  counter              852950 non-null  int64         
 14  year_occurred        852950 non-null  int64         
 15  month_occurred       852950 non-null  int64         
 16  hour_occurred        852950 non-null  int64         
dtypes: datetime64[ns](1), float64(2), int64(5), object(9)
memory usage: 110.6+ MB

null_counts = df.isnull().sum()
# Convert to DataFrame for better visualization
null_counts_df = pd.DataFrame(null_counts, columns=['Number of Nulls'])
null_counts_df = null_counts_df[null_counts_df['Number of Nulls']>0]
null_counts_df['Nulls_share'] = null_counts_df['Number of Nulls'] / len(df)
null_counts_df['Nulls_share'] =null_counts_df['Nulls_share'] .round(2)
null_counts_df =null_counts_df.sort_values(by='Nulls_share',ascending=False)
null_counts_df

	Number of Nulls	Nulls_share
weapon_description	556202	0.65
victim_sex	112606	0.13
victim_descent	112614	0.13
premise_description	518	0.00

df.describe()

	date_occurred	victim_age	latitude	longitude	counter	year_occurred	month_occurred	hour_occurred
count	852950	852950.000000	852950.000000	852950.000000	852950.0	852950.000000	852950.000000	852950.000000
mean	2022-01-09 05:17:37.460835840	29.742191	33.983232	-118.040106	1.0	2021.532406	6.419488	13.185609
min	2020-01-01 00:01:00	-3.000000	0.000000	-118.667600	1.0	2020.000000	1.000000	0.000000
25%	2021-01-26 15:00:00	5.000000	34.014100	-118.429700	1.0	2021.000000	4.000000	9.000000
50%	2022-01-30 12:42:00	31.000000	34.058500	-118.321500	1.0	2022.000000	6.000000	14.000000
75%	2022-12-25 15:00:00	45.000000	34.163200	-118.273900	1.0	2022.000000	9.000000	19.000000
max	2023-12-04 23:00:00	120.000000	34.334300	0.000000	1.0	2023.000000	12.000000	23.000000
std	NaN	21.799470	1.756263	6.089068	0.0	1.099147	3.380271	6.535645

Crimes & victimes

Crime description

df.groupby('crime_description')['counter'].count().reset_index().sort_values(by='counter',ascending=False).head(5)

	crime_description	counter
132	VEHICLE - STOLEN	91473
4	BATTERY - SIMPLE ASSAULT	67976
115	THEFT OF IDENTITY	53467
21	BURGLARY FROM VEHICLE	52611
20	BURGLARY	51961

Premise description

df.groupby('premise_description')['counter'].count().reset_index().sort_values(by='counter',ascending=False).head(10)

	premise_description	counter
266	STREET	216018
255	SINGLE FAMILY DWELLING	144367
207	MULTI-UNIT DWELLING (APARTMENT, DUPLEX, ETC)	104461
227	PARKING LOT	59551
218	OTHER BUSINESS	40412
254	SIDEWALK	36700
298	VEHICLE, PASSENGER/TRUCK	25224
79	GARAGE/CARPORT	16673
63	DRIVEWAY	13838
246	RESTAURANT/FAST FOOD	10841

Weapon description

#weapon_description

df.groupby('weapon_description')['counter'].count().reset_index().sort_values(by='counter',ascending=False).head(10)

	weapon_description	counter
65	STRONG-ARM (HANDS, FIST, FEET OR BODILY FORCE)	159021
75	UNKNOWN WEAPON/OTHER WEAPON	31728
78	VERBAL THREAT	21767
27	HAND GUN	18330
58	SEMI-AUTOMATIC PISTOL	6657
32	KNIFE WITH BLADE 6INCHES OR LESS	6262
73	UNKNOWN FIREARM	6011
44	OTHER KNIFE	5361
39	MACE/PEPPER SPRAY	3390
77	VEHICLE	2979

Status description

#weapon_description

df.groupby('status_description')['counter'].count().reset_index().sort_values(by='counter',ascending=False).head(10)

	status_description	counter
2	Invest Cont	683107
1	Adult Other	91576
0	Adult Arrest	74004
3	Juv Arrest	2775
4	Juv Other	1484
5	UNK	4

Place

df.groupby('area_name')['counter'].count().reset_index().sort_values(by='counter',ascending=False).head(10)

	area_name	counter
1	Central	57691
0	77th Street	53631
12	Pacific	49850
15	Southwest	47858
6	Hollywood	45018
14	Southeast	43357
11	Olympic	42979
9	Newton	42743
8	N Hollywood	42455
20	Wilshire	40648

Crimes per place and years evolution

df.groupby(['area_name','year_occurred'])['counter'].count().reset_index().sort_values(by='counter',ascending=False).head(10)

	area_name	year_occurred	counter
6	Central	2022	17673
7	Central	2023	15273
2	77th Street	2022	14532
62	Southwest	2022	13371
0	77th Street	2020	13309
5	Central	2021	13156
1	77th Street	2021	13074
50	Pacific	2022	13013
49	Pacific	2021	12917
3	77th Street	2023	12716

Crime evolution per place

df_evo = df.query(" year_occurred in (2021, 2023)  ").pivot_table(index='area_name', columns='year_occurred',aggfunc='sum',values='counter').reset_index()
df_evo.columns=['area_name','2021','2023']
df_evo.head()

	area_name	2021	2023
0	77th Street	13074	12716
1	Central	13156	15273
2	Devonshire	8487	8665
3	Foothill	7032	6418
4	Harbor	8858	8212

import pandas as pd
import plotly.express as px

# Assuming df_evo is your DataFrame and it's already defined

# Calculate the evolution from 2021 to 2023
df_evo['evol'] = df_evo['2023'].astype(float) / df_evo['2021'].astype(float) - 1
df_evo['evol'] = df_evo['evol'].round(2)

# Sort the DataFrame based on the evolution
df_evo = df_evo.sort_values(by='evol')

# Categorize the evolution into colors
df_evo['colors'] = df_evo['evol'].apply(lambda x:
                                        'red' if x > 0.05
                                        else ('green' if x < -0.05
                                        else 'grey'))

# Define specific colors for each category
color_map = {
    'green': '#32CD32',  # Light green for positive evolution
    'red': '#FF6347',    # Tomato red for negative evolution
    'grey': '#808080'    # Grey for neutral
}

# Create a scatter plot with specified colors and area names
fig = px.scatter(df_evo, x="evol", y="2023", color='colors', title='Evolution of 2023 Compared to 2021',
                 color_discrete_map=color_map,
                 text='area_name')  # Use 'area_name' instead of 'areaname'

# Set all markers to have grey borders and 70% opacity
fig.update_traces(marker=dict(size=12, line=dict(width=1, color='grey'), opacity=0.7))

# Improve layout for better readability
fig.update_traces(textposition='top center')
fig.update_layout(
    showlegend=False,
    plot_bgcolor='white',  # Set plot background to white
    paper_bgcolor='white'  # Set overall figure background to white
)

# Show the plot
fig.show('notebook')

Central seems to be the most dangerous place, where crimes are also growing a lot.

fig = px.line(df.groupby(['area_name','year_occurred'])['counter'].count().reset_index().sort_values(by=['area_name','year_occurred'],ascending=False)
              , x="year_occurred"
                , y="counter"
                ,color = 'area_name'
                , title='Crime evolution per place'
                ,markers=True
                )
fig.show('notebook')

Victim age

df.groupby('victim_age')['counter'].count().reset_index().sort_values(by='counter',ascending=False).head(5)

	victim_age	counter
3	0	211842
32	30	19421
37	35	19008
33	31	18603
31	29	18552

import plotly.express as px

# Assuming df is your DataFrame and it's already defined
fig = px.histogram(df, x="victim_age", title='Distribution of victims per age')

# Calculate the number of victims with age 0
num_victims_no_age = len(df[df['victim_age'] == 0])

# Add an annotation with an arrow
fig.add_annotation(
    x=0,  # Specifies the x-coordinate of the arrow's head
    y=max(df['victim_age'].value_counts()),  # Adjust this value as needed for y-coordinate
    text=f"{num_victims_no_age} victims don't have an attributed age",  # The text of the annotation
    showarrow=True,  # Show an arrow pointing to the specified coordinates
    arrowhead=1  # The style of the arrowhead (1 is a simple arrow)
)

# Show the figure
fig.show('notebook')

# Assuming df is your DataFrame and it's already defined
fig = px.histogram( df.query(" victim_sex=='F' ")
                   , x="victim_age"
                   , color = 'victim_sex'
                   ,title='Distribution of female victims per age')



# Show the figure
fig.show('notebook')

# Assuming df is your DataFrame and it's already defined
fig = px.histogram( df.query(" victim_sex=='M' ")
                   , x="victim_age"
                   , color = 'victim_sex'
                   ,title='Distribution of males victims per age')



# Show the figure
fig.show('notebook')

Victim sex

df.groupby('victim_sex')['counter'].count().reset_index().sort_values(by='counter',ascending=False).head(5)

	victim_sex	counter
3	M	351362
1	F	313468
4	X	75420
2	H	93
0	-	1

Victim descent

df.groupby('victim_descent')['counter'].count().reset_index().sort_values(by='counter',ascending=False).head(5)

	victim_descent	counter
7	H	261145
17	W	173440
2	B	120896
18	X	83214
12	O	67532

Victims sex / descent

df_victims = df.pivot_table( index='victim_descent',columns='victim_sex',aggfunc='sum',values='counter').reset_index()
df_victims = df_victims[['victim_descent','M','F']]
df_victims=df_victims.fillna(0)
df_victims['total']=df_victims['M']+df_victims['F']
df_victims = df_victims.sort_values(by='total',ascending=False)

df_victims['female_share'] = df_victims['F']  / df_victims['total']
df_victims['female_share']=df_victims['female_share'].round(2)

df_victims.head(5)

victim_sex	victim_descent	M	F	total	female_share
7	H	129647.0	131318.0	260965.0	0.50
17	W	100225.0	72995.0	173220.0	0.42
2	B	53081.0	67676.0	120757.0	0.56
12	O	43695.0	23741.0	67436.0	0.35
1	A	9657.0	9034.0	18691.0	0.48

Crimes by dates

df.head(1)

	date_reported	date_occurred	area_name	crime_description	victim_age	victim_sex	victim_descent	premise_description	weapon_description	status_description	location	latitude	longitude	counter	year_occurred	month_occurred	hour_occurred
0	2020-01-08	2020-01-08 22:30:00	Southwest	BATTERY - SIMPLE ASSAULT	36	F	B	SINGLE FAMILY DWELLING	STRONG-ARM (HANDS, FIST, FEET OR BODILY FORCE)	Adult Other	1100 W 39TH PL	34.0141	-118.2978	1	2020	1	22

fig = px.line(df.pivot_table(index='year_occurred' ,aggfunc='sum' , values='counter'  ).reset_index()
                , x="year_occurred"
                , y="counter"
                , title='Nb of crimes per year'
                ,markers=True
                )
fig.show('notebook')

df_grouped = df.pivot_table(index=['month_occurred'] ,aggfunc='sum' , values='counter'  ).reset_index()
df_grouped['counter_mean']=df_grouped['counter'] / len(df_grouped)
df_grouped

	month_occurred	counter	counter_mean
0	1	73129	6094.083333
1	2	68679	5723.250000
2	3	71196	5933.000000
3	4	70299	5858.250000
4	5	73349	6112.416667
5	6	72891	6074.250000
6	7	75429	6285.750000
7	8	75072	6256.000000
8	9	72141	6011.750000
9	10	75464	6288.666667
10	11	69960	5830.000000
11	12	55341	4611.750000

fig = px.line(df_grouped
                , x="month_occurred"
                , y="counter_mean"
                #,color ='year_occurred'
                ,title='Crime Month Seasonality'
                ,markers=True
                )
fig.show('notebook')

fig = px.line(df.pivot_table(index=['year_occurred','month_occurred'] ,aggfunc='sum' , values='counter'  ).reset_index()
                , x="month_occurred"
                , y="counter"
                ,color ='year_occurred'
                ,title='Crime Month Seasonality'
                ,markers=True
                )
fig.show('notebook')

import plotly.express as px

# Assuming df is your DataFrame and it's already defined
fig = px.line(df.pivot_table(index=['year_occurred', 'month_occurred'], aggfunc='sum', values='counter').reset_index(),
              x="month_occurred",
              y="counter",
              color='year_occurred',
              title='Crime Month Seasonality',
              markers=True)

# Add an annotation with an arrow for December 2023
fig.add_annotation(
    x=12,  # x-coordinate for December
    y=df[(df['year_occurred'] == 2023) & (df['month_occurred'] == 12)]['counter'].sum(),  # y-coordinate for December 2023
    text="Missing data for December 2023",  # The annotation text
    showarrow=True,  # Show an arrow pointing to the specified coordinates
    arrowhead=1,  # The style of the arrowhead (1 is a simple arrow)
    ax=-120,  # Horizontal offset of the arrow (adjust as needed)
    ay=-50  # Vertical offset of the arrow (adjust as needed)
)

# Show the figure
fig.show('notebook')

Crimes correlations : What crimes correlate with each - others ?

# Assuming you have a DataFrame 'df' with columns 'area_name', 'counter', and 'crime_description'
df_crimes = df.pivot_table(index='area_name', aggfunc='sum', values='counter', columns='crime_description').reset_index()

# Set 'area_name' as the index
df_crimes = df_crimes.set_index('area_name')

import pandas as pd

# Assuming df_crimes is your DataFrame
# df_crimes = pd.read_csv('your_data.csv')  # Example to load data

# Calculate the correlation matrix
correlation_matrix = df_crimes.corr()

# Display the correlation matrix
correlation_matrix

crime_description	ARSON	ASSAULT WITH DEADLY WEAPON ON POLICE OFFICER	ASSAULT WITH DEADLY WEAPON, AGGRAVATED ASSAULT	ATTEMPTED ROBBERY	BATTERY - SIMPLE ASSAULT	BATTERY ON A FIREFIGHTER	BATTERY POLICE (SIMPLE)	BATTERY WITH SEXUAL CONTACT	BEASTIALITY, CRIME AGAINST NATURE SEXUAL ASSLT WITH ANIM	BIGAMY	...	UNAUTHORIZED COMPUTER ACCESS	VANDALISM - FELONY ($400 & OVER, ALL CHURCH VANDALISMS)	VANDALISM - MISDEAMEANOR ($399 OR UNDER)	VEHICLE - ATTEMPT STOLEN	VEHICLE - STOLEN	VEHICLE, STOLEN - OTHER (MOTORIZED SCOOTERS, BIKES, ETC)	VIOLATION OF COURT ORDER	VIOLATION OF RESTRAINING ORDER	VIOLATION OF TEMPORARY RESTRAINING ORDER	WEAPONS POSSESSION/BOMBING
crime_description
ARSON	1.000000	0.640382	0.802754	0.877036	0.795099	0.414322	0.792373	0.416257	0.319015	-0.408223	...	-0.575377	0.762680	0.785109	0.394897	0.627737	0.012203	-0.448538	-0.021869	-0.582394	0.220270
ASSAULT WITH DEADLY WEAPON ON POLICE OFFICER	0.640382	1.000000	0.558682	0.599608	0.674356	0.542911	0.737177	0.470088	0.203734	-0.227827	...	-0.452161	0.585272	0.516159	0.114434	0.181087	0.079398	-0.223461	-0.031021	-0.356584	0.038231
ASSAULT WITH DEADLY WEAPON, AGGRAVATED ASSAULT	0.802754	0.558682	1.000000	0.921664	0.722539	0.267665	0.812042	0.191782	0.053449	-0.492927	...	-0.614649	0.488878	0.875467	0.604097	0.764485	0.004481	-0.276640	0.272391	-0.298845	0.282560
ATTEMPTED ROBBERY	0.877036	0.599608	0.921664	1.000000	0.879820	0.461832	0.827545	0.440890	0.361852	-0.538950	...	-0.534217	0.678861	0.807360	0.424522	0.608391	0.014359	-0.383709	0.117981	-0.351585	0.293896
BATTERY - SIMPLE ASSAULT	0.795099	0.674356	0.722539	0.879820	1.000000	0.666625	0.817322	0.749157	0.443872	-0.418397	...	-0.397360	0.864910	0.695500	0.250183	0.315457	0.182661	-0.527686	-0.083608	-0.382659	0.250179
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
VEHICLE, STOLEN - OTHER (MOTORIZED SCOOTERS, BIKES, ETC)	0.012203	0.079398	0.004481	0.014359	0.182661	0.201154	0.125646	0.360140	0.509747	0.983821	...	-0.029351	0.331106	0.237177	0.437324	0.112618	1.000000	-0.078017	-0.093430	-0.192871	0.273810
VIOLATION OF COURT ORDER	-0.448538	-0.223461	-0.276640	-0.383709	-0.527686	-0.322065	-0.422450	-0.534585	0.017299	0.308828	...	0.036071	-0.579071	-0.269040	0.042368	-0.014250	-0.078017	1.000000	0.278040	0.416653	-0.028942
VIOLATION OF RESTRAINING ORDER	-0.021869	-0.031021	0.272391	0.117981	-0.083608	-0.261454	-0.058197	-0.435435	-0.625504	-0.146899	...	-0.496845	-0.287178	0.215279	0.412006	0.519665	-0.093430	0.278040	1.000000	-0.140528	0.325251
VIOLATION OF TEMPORARY RESTRAINING ORDER	-0.582394	-0.356584	-0.298845	-0.351585	-0.382659	-0.197052	-0.416955	-0.239865	0.363688	-0.571571	...	0.407601	-0.587476	-0.521707	-0.235402	-0.325704	-0.192871	0.416653	-0.140528	1.000000	-0.349798
WEAPONS POSSESSION/BOMBING	0.220270	0.038231	0.282560	0.293896	0.250179	0.212912	0.356509	0.195228	-0.722222	0.547723	...	-0.358772	0.308248	0.275263	0.256395	0.252524	0.273810	-0.028942	0.325251	-0.349798	1.000000

138 rows × 138 columns

Clusters

import pandas as pd
from sklearn.preprocessing import StandardScaler
from sklearn.cluster import KMeans

# Assuming you have a DataFrame 'df' with columns 'area_name', 'counter', and 'crime_description'
# Creating the df_crimes DataFrame
df_crimes = df.pivot_table(index='area_name', aggfunc='sum', values='counter', columns='crime_description')
df_crimes = df_crimes.fillna(0)

# Preparing the data for clustering
X = df_crimes.values  # Extracting the numerical values for clustering

# Feature Scaling
scaler = StandardScaler()
X_scaled = scaler.fit_transform(X)

# Applying K-Means Clustering
k = 5  # Number of clusters
kmeans = KMeans(n_clusters=k, random_state=42)
kmeans.fit(X_scaled)

# Attaching the cluster labels to the original DataFrame
df_crimes['Cluster'] = kmeans.labels_

# Displaying the first few rows of the DataFrame with cluster labels
df_crimes.sort_values(by='Cluster')

C:\Users\antoi\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.11_qbz5n2kfra8p0\LocalCache\local-packages\Python311\site-packages\sklearn\cluster\_kmeans.py:1416: FutureWarning:

The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning

crime_description	ARSON	ASSAULT WITH DEADLY WEAPON ON POLICE OFFICER	ASSAULT WITH DEADLY WEAPON, AGGRAVATED ASSAULT	ATTEMPTED ROBBERY	BATTERY - SIMPLE ASSAULT	BATTERY ON A FIREFIGHTER	BATTERY POLICE (SIMPLE)	BATTERY WITH SEXUAL CONTACT	BEASTIALITY, CRIME AGAINST NATURE SEXUAL ASSLT WITH ANIM	BIGAMY	...	VANDALISM - FELONY ($400 & OVER, ALL CHURCH VANDALISMS)	VANDALISM - MISDEAMEANOR ($399 OR UNDER)	VEHICLE - ATTEMPT STOLEN	VEHICLE - STOLEN	VEHICLE, STOLEN - OTHER (MOTORIZED SCOOTERS, BIKES, ETC)	VIOLATION OF COURT ORDER	VIOLATION OF RESTRAINING ORDER	VIOLATION OF TEMPORARY RESTRAINING ORDER	WEAPONS POSSESSION/BOMBING	Cluster
area_name
Foothill	86.0	52.0	1562.0	93.0	2173.0	6.0	60.0	126.0	1.0	0.0	...	1705.0	789.0	71.0	3585.0	7.0	370.0	474.0	44.0	2.0	0
Harbor	84.0	23.0	2104.0	153.0	2815.0	9.0	74.0	111.0	0.0	0.0	...	2294.0	1182.0	235.0	4885.0	26.0	353.0	1096.0	17.0	3.0	0
Hollenbeck	95.0	59.0	2203.0	156.0	2885.0	10.0	80.0	116.0	0.0	0.0	...	2092.0	1119.0	197.0	5029.0	3.0	368.0	694.0	45.0	1.0	0
Mission	92.0	35.0	1798.0	149.0	2297.0	6.0	78.0	177.0	0.0	0.0	...	1960.0	801.0	108.0	4662.0	2.0	338.0	925.0	35.0	0.0	0
Newton	172.0	49.0	3503.0	339.0	3904.0	6.0	97.0	165.0	0.0	1.0	...	2571.0	1159.0	181.0	6693.0	5.0	208.0	831.0	38.0	3.0	0
Southwest	104.0	77.0	3197.0	272.0	4065.0	7.0	127.0	218.0	0.0	0.0	...	2787.0	1415.0	255.0	5204.0	546.0	331.0	743.0	45.0	3.0	0
Rampart	125.0	37.0	2883.0	329.0	3833.0	19.0	93.0	195.0	0.0	1.0	...	2437.0	1128.0	179.0	4366.0	58.0	253.0	438.0	36.0	2.0	0
Central	199.0	151.0	3755.0	436.0	6276.0	41.0	300.0	369.0	2.0	1.0	...	4029.0	1389.0	115.0	3755.0	118.0	133.0	260.0	9.0	2.0	1
77th Street	183.0	71.0	5334.0	447.0	4195.0	7.0	258.0	158.0	0.0	0.0	...	2893.0	1602.0	238.0	7046.0	3.0	218.0	589.0	16.0	2.0	2
Southeast	149.0	82.0	4399.0	321.0	3500.0	10.0	202.0	116.0	1.0	0.0	...	2255.0	1592.0	282.0	5825.0	1.0	264.0	690.0	36.0	0.0	2
West LA	73.0	5.0	722.0	94.0	2648.0	7.0	37.0	198.0	0.0	0.0	...	2485.0	923.0	48.0	2831.0	162.0	163.0	297.0	17.0	1.0	3
Van Nuys	87.0	12.0	1467.0	107.0	2569.0	3.0	70.0	162.0	1.0	0.0	...	2309.0	932.0	84.0	3636.0	32.0	343.0	385.0	33.0	0.0	3
Topanga	45.0	29.0	1417.0	112.0	2496.0	13.0	55.0	159.0	0.0	0.0	...	1803.0	676.0	84.0	2704.0	16.0	215.0	652.0	89.0	1.0	3
Northeast	119.0	30.0	1541.0	185.0	2419.0	10.0	84.0	114.0	0.0	0.0	...	2471.0	1048.0	140.0	4337.0	25.0	384.0	537.0	24.0	3.0	3
West Valley	75.0	28.0	1596.0	136.0	2457.0	14.0	64.0	179.0	2.0	0.0	...	2112.0	819.0	117.0	3393.0	19.0	463.0	173.0	130.0	1.0	3
N Hollywood	127.0	32.0	1763.0	141.0	3036.0	5.0	85.0	190.0	0.0	1.0	...	2816.0	1061.0	141.0	3960.0	48.0	87.0	102.0	26.0	1.0	3
Devonshire	47.0	13.0	1192.0	101.0	2361.0	4.0	56.0	147.0	0.0	0.0	...	1716.0	671.0	77.0	3151.0	34.0	315.0	339.0	131.0	0.0	3
Wilshire	119.0	118.0	1550.0	160.0	2898.0	7.0	100.0	174.0	0.0	0.0	...	2551.0	945.0	54.0	3326.0	47.0	307.0	354.0	11.0	1.0	3
Olympic	133.0	25.0	2470.0	280.0	4053.0	6.0	84.0	205.0	2.0	1.0	...	2671.0	1186.0	146.0	4652.0	46.0	293.0	447.0	42.0	1.0	4
Hollywood	116.0	54.0	2540.0	270.0	4102.0	19.0	189.0	270.0	1.0	0.0	...	3052.0	1022.0	62.0	3211.0	53.0	194.0	396.0	21.0	5.0	4
Pacific	119.0	32.0	1880.0	154.0	2994.0	20.0	139.0	216.0	0.0	2.0	...	2817.0	1131.0	296.0	5222.0	544.0	255.0	327.0	10.0	3.0	4

21 rows × 139 columns

df_evo_2 = df_evo.merge(df_crimes,on='area_name')
df_evo_2.head(1)

	area_name	2021	2023	evol	colors	ARSON	ASSAULT WITH DEADLY WEAPON ON POLICE OFFICER	ASSAULT WITH DEADLY WEAPON, AGGRAVATED ASSAULT	ATTEMPTED ROBBERY	BATTERY - SIMPLE ASSAULT	...	VANDALISM - FELONY ($400 & OVER, ALL CHURCH VANDALISMS)	VANDALISM - MISDEAMEANOR ($399 OR UNDER)	VEHICLE - ATTEMPT STOLEN	VEHICLE - STOLEN	VEHICLE, STOLEN - OTHER (MOTORIZED SCOOTERS, BIKES, ETC)	VIOLATION OF COURT ORDER	VIOLATION OF RESTRAINING ORDER	VIOLATION OF TEMPORARY RESTRAINING ORDER	WEAPONS POSSESSION/BOMBING	Cluster
0	Hollywood	12084	10274	-0.15	green	116.0	54.0	2540.0	270.0	4102.0	...	3052.0	1022.0	62.0	3211.0	53.0	194.0	396.0	21.0	5.0	4

1 rows × 144 columns

df_evo_2['Cluster']=df_evo_2['Cluster'].astype(str)
fig = px.scatter(df_evo_2
                 , x="evol", y="2023", color="Cluster",
                 size='2023', hover_data=['2023'])
fig.show('notebook')

import pandas as pd
import plotly.express as px

# Ensure df_evo_2 is correctly set up with 'evol', '2023', 'Cluster', and 'area_name' columns

# Convert Cluster to a string for color coding
df_evo_2['Cluster'] = df_evo_2['Cluster'].astype(str)

# Create the scatter plot
fig = px.scatter(df_evo_2, x="evol", y="2023", color="Cluster",
                 size='2023', hover_data=['area_name', '2023'],
                 text='area_name')  # Add area name as text labels on the plot

# Customize the layout
fig.update_traces(textposition='top center')
fig.update_layout(
    title="Evolution of 2023 Compared to 2021 by Cluster",
    xaxis_title="Evolution from 2021 to 2023",
    yaxis_title="Value in 2023",
    legend_title="Cluster",
    plot_bgcolor='white',
    paper_bgcolor='white'
)

# Customize hover template for better readability
fig.update_traces(hovertemplate="<br>".join([
    "Area: %{hoverdata[0]}",
    "2023 Value: %{hoverdata[1]}",
    "Evolution: %{x}",
    "Cluster: %{marker.color}"
]))

# Show the plot
fig.show('notebook')

THE END